In conventional storage systems, the storage system typically can receive a plurality of read/write requests. The target area for the executed read/write request will be locked, thereby avoiding conflicts between a plurality of read/write requests. Meanwhile, the total target area of the read/write requests which are suspended due to partial lock of the target area will also be locked. Therefore, the storage space which is in fact not accessed cannot be read and written due to the lockup, thereby decreasing the overall response speed of the storage system.
Thus, it has become the focus of attention that how to improve concurrency level of a plurality of read/write requests for the storage system without any read/write request conflict.